CN114641925A

CN114641925A - Safe active discharge circuit for inverter in vehicle

Info

Publication number: CN114641925A
Application number: CN202080075667.1A
Authority: CN
Inventors: 亨里克·古尔文; 罗尔夫·安德森; 托尔比约恩·西蒙森
Original assignee: Aros Electronics AB
Current assignee: Aros Electronics AB
Priority date: 2019-11-08
Filing date: 2020-11-06
Publication date: 2022-06-17
Also published as: EP4055704A1; EP3820030A1; US20220393571A1; EP4055704B1; US11955882B2; WO2021089831A1

Abstract

An active discharge circuit for an electric vehicle inverter intended to be connected in parallel with a DC link capacitor connected between positive and negative lines of a DC power link, wherein the circuit comprises a dissipation current source, a switch connected in series with the current source between the DC lines, and a controller connected to the switch and arranged to apply an activation signal according to a control signal, the activation signal placing the switch in a conducting state, wherein the current source is configured to draw a discharge current and dissipate any energy stored in the DC link capacitor when the switch is in the conducting state. As long as the switch is closed by the activation signal, the current source draws a constant current and dissipates power and the voltage across the DC-link capacitor decreases linearly.

Description

Safe active discharge circuit for inverter in vehicle

Technical Field

The present invention relates to a safety active discharge circuit arranged in parallel with a DC link capacitor connected between the positive and negative lines of a DC power link.

Background

Inverters in vehicles (and also in many other applications) receive a high voltage input and provide an alternating current to drive, for example, AC machines. The switches of the inverter need to operate at high frequencies and a stable and reliable DC input voltage is needed. For this reason, a capacitor (referred to as a DC link capacitor) is generally connected between the positive and negative lines to absorb a ripple caused by switching in the inverter. Thus, the DC-link capacitor ensures a stable and reliable voltage across the inverter.

The high voltage input voltage is received through a DC link, which is basically two power lines connected to a DC power source (e.g. a high voltage battery) via mechanical circuit breakers (relays). These circuit breakers enable the inverter to be quickly disconnected from the battery in the event of a shutdown, for example, caused by a key shutdown, a power failure, or a vehicle collision. However, the DC link capacitor will still be charged and this charge needs to be discharged for safety reasons.

A traditional straightforward solution is to hard-wire a passive resistor across the capacitor. The resistance is high enough to prevent excessive power loss during normal operation. Therefore, the discharge through this resistor will take a relatively long time, typically on the order of a few minutes. Under normal circumstances, such as key-off, this is not usually a problem. However, in some cases, such as collisions, there are on-site safety regulations that require a faster discharge process, for example, on the order of five seconds.

To provide such a fast discharge, a much lower discharge resistance in series with the switch may be connected across the capacitor. The switch is connected to disconnect the resistor whenever the switch receives a disable command from the control unit. If a fault occurs, the disable command will no longer be present and the switch will connect the discharge resistor, causing the capacitor to discharge quickly. Such a switch-controlled discharge resistor is referred to as an active discharge resistor.

One challenge with such an active discharge resistor is to cope with the situation where the capacitor is continuously charged. For example, if the circuit breaker fails to open, the battery may still be connected, or, in the case of vehicle travel, the AC motor may generate a back emf that is coupled to the capacitor. Such a continuous voltage supply will cause the active discharge resistor to discharge the continuous energy supply. Thus, conventional active discharge resistors require sufficient power capacity and thermal rating for continuous operation, which results in a significant increase in cost.

US2017/0355267 proposes a solution to this problem, wherein a timing circuit is introduced to further control the switching of the active discharge resistor. According to this solution, the timing circuit will disconnect the discharge resistor after a predetermined period of time if there is still voltage across the capacitor. However, this solution is relatively complex and expensive. EP 3468019 provides another example.

There is a need for an improved active discharge circuit that avoids the above-mentioned disadvantages.

General disclosure of the invention

According to a first aspect of the invention, this and other objects are achieved by an active discharge circuit for an electric vehicle inverter intended to be connected in parallel with a DC link capacitor connected between positive and negative lines of a DC power link and configured to discharge said DC link capacitor in less than 7 seconds, wherein the circuit comprises a dissipation current source, a switch connected in series with the current source between the DC lines, and a controller connected to the switch and arranged to apply an activation signal according to a control signal from a vehicle control system, the activation signal placing the switch in a conducting state, wherein the current source is configured to draw a discharge current and dissipate any energy stored in said DC link capacitor when the switch is in the conducting state.

With this design, the energy in the capacitor will be dissipated by the current source rather than the passive resistor. As long as the switch is closed by the activation signal, the current source draws a constant current and dissipates power and the voltage across the DC-link capacitor decreases linearly.

An active discharge circuit using a dissipative current source according to the invention has improved performance at similar power ratings or, in other words, requires a lower power rating to meet performance requirements. Here, performance relates to discharging a given high voltage below a given limit in a given time.

As an example, a typical prior art active discharge resistor may have a power rating of 24W, while the active component of the present invention may have a power rating of only 7W as a performance contrast.

In a preferred embodiment, the current source is connected in series with one or several additional active components. Each active component will provide additional power dissipation capability. This has the advantage of increasing the thermal mass at low cost (many small parts rather than one large part).

The active components may be transistors connected in a source-drain chain. The transistors will operate in the linear (non-saturated) region, i.e. each transistor has a current through it and a voltage across it. The transistor may be, for example, a Field Effect Transistor (FET), an Insulated Gate Bipolar Transistor (IGBT), or a Bipolar Junction Transistor (BJT).

The circuit may further comprise a set of resistors connected in series across the DC-link capacitor to divide the voltage across the DC-link capacitor into a set of intermediate voltages, each intermediate voltage being connected to the gate of one of the (field effect) transistors. Thus, each transistor will be turned on as long as there is a voltage across the DC link capacitor.

In one embodiment, the dissipation current source includes a transistor and a voltage regulator connected between the gate of the field effect transistor and the negative DC line. A constant voltage is applied to the gate of the transistor as long as the bias current flows through the voltage regulator. This will result in a constant current (discharge current) flowing through the transistor when the switch is placed in the on state.

Also here, the transistor may be, for example, a Field Effect Transistor (FET), an Insulated Gate Bipolar Transistor (IGBT), or a Bipolar Junction Transistor (BJT). The voltage regulator may be implemented by a zener diode, a transient voltage suppression diode, or a voltage reference IC.

Preferably, the drain of the transistor is connected to the positive line without any intermediate resistive load. This means that the discharge current does not cause any resistive dissipation, i.e. all dissipation is provided by the active circuit (i.e. the current source and any additional active components).

In some embodiments, a predefined "idle" current is also allowed to flow through the dissipative element (transistor) when the switch is in a non-conductive state. Such a reactive current is significantly smaller than the discharge current. For example, the current may be less than 1mA, such as less than 0.5mA or less than 0.1mA, while the discharge current may be approximately 5-50 mA. This reactive current can be used to power the controller, thereby providing a safe independent power source whenever the link capacitor is charging.

In one embodiment, the switch comprises a field effect transistor having a drain connected to the set of active components, a source connected to ground (via a resistor), and a gate connected to receive the activation signal.

The controller may be configured to apply a steady activation signal to control the current source to draw a constant discharge current such that the voltage across the DC link capacitor drops linearly. Optionally, the controller is configured to apply an intermittent activation signal, thereby allowing a non-constant discharge current. For example, the current source may be controlled to draw an increasing current in order to dissipate a constant power, such that the energy discharge of the DC link capacitor is linear. Alternatively, the current source may be controlled to draw a decreasing current such that the voltage across the DC link capacitor decreases exponentially. This effectively corresponds to the performance of the passive discharge resistor, which may be advantageous if the discharge should be synchronized with other discharge processes.

The control signals may be communicated over a bi-directional serial communication bus. Such a serial communication bus is typically already present in a vehicle and provides a simple way of accessing the controller. To ensure the required security, communication may be provided over a "black channel", for example involving a specific (secure) communication protocol.

Optionally, the control signal is a (unidirectional) discharge request signal, and the controller is configured to apply the activation signal when there is no discharge request signal. Such a signal is always present during normal operation of the vehicle, but in the event of a fault, the discharge request signal is no longer provided and an active discharge should be activated.

In one embodiment, the active discharge circuit includes circuitry for pulsing the discharge request signal to generate a pulsed discharge signal, and the controller is configured to verify the pulsed discharge signal and provide the activation signal when verification is unsuccessful. For example, the controller may be configured to verify a pulse width, pattern, and/or pulse repetition frequency of the pulsed discharge signal. This is done to ensure signal integrity and to detect potential faults in the input circuit.

It is noted that such pulsing of the control signal may be considered to represent another inventive concept in order to make the control signal more reliable, which may also be beneficial in other situations than the active discharge circuit according to the first aspect of the invention described above. Indeed, in any case where no control signal is used for a system fault, modulation (pulsing or other modulation) of that control signal using the system's inherent power supply will enable verification of the complete signal path of the control signal, i.e. all components between the control signal input to the controller. For example, an interface between a high voltage domain and a low voltage domain, such as an optocoupler, may fail, resulting in a constant "high" level. By pulsing the control signal, such a fault is immediately detected. Furthermore, in the event that the control signal remains "high" even during a system fault, the system fault will also typically result in a loss of power to the low voltage circuit, and thus a modulation interruption. Thus, a system failure will still be detected.

The switch may be connected to the negative line via a resistor, and the controller may be connected to detect the voltage across the resistor. The detected voltage is indicative of the current through the resistor and can be used for simple functional testing. To ensure functional availability, a short (millisecond) activation signal may be applied when measuring the current generated. The controller can thus verify the correct operation of the active discharge circuit.

According to one embodiment, the active discharge circuit may further comprise a voltmeter connected to detect a link voltage between the DC lines (i.e., across the link capacitor), and the controller may then be connected to receive an indication of the link voltage from the voltmeter and determine whether the link voltage has dropped correctly, and to cause the switch to enter the non-conductive state when it is determined that the link voltage has not dropped correctly.

This allows the controller to immediately disable the active discharge process, preventing the voltage from decreasing as expected in the event of a fault condition. This may occur, for example, if the DC power source is not properly disconnected from the inverter for some reason. By disabling the active discharge, thermal events (overheating, etc.) in the discharge circuit can be avoided. The active discharge circuit according to this embodiment of the invention may comply with relevant safety regulations for electric vehicles, such as ISO 26262 and IEC 13849.

Brief Description of Drawings

The present invention will be described in more detail with reference to the appended drawings, which illustrate currently preferred embodiments of the invention.

Fig. 1 is a schematic block diagram of an active discharge circuit connected across a DC link capacitor of an electric vehicle inverter, in accordance with an embodiment of the present invention.

Fig. 2 is a more detailed circuit diagram of an embodiment of the active discharge circuit of fig. 1.

Fig. 3 is a circuit diagram of a discharge signal processing circuit according to an embodiment of the present invention.

Detailed description of the presently preferred embodiments

Fig. 1 shows an inverter circuit 1 connected to a DC power source 2 via a DC link comprising a positive line 3 and a negative line 4. A DC link capacitor 5 is connected in parallel with the inverter 1 and a high resistance passive discharge resistor 6 is connected in parallel with the link capacitor to ensure discharge of the capacitor in the event of inverter failure or power loss.

The inverter is here connected to provide an AC voltage for the electric machine in the electric vehicle. The electric motor may form part of a vehicle traction system, but may alternatively be an electric motor for some other device, such as power regeneration, an air compressor, a water pump, etc. For example, the voltage across the

lines

3, 4 of the DC link is 800V.

An active discharge circuit 10 according to an embodiment of the present invention is also connected in parallel with the link capacitor 5. Here, the active discharge circuit 10 comprises a current source 11 connected in series with a switch 12 across the link capacitor 5. The controller 13 is connected to control the switch 12. Furthermore, in the embodiment shown, an additional dissipation element 14 is connected in series with the current source 11.

The active discharge circuit 10 may also include a voltmeter 15, the voltmeter 15 being connected to measure the voltage across the DC link and provide a signal indicative thereof to the controller 13.

The controller 13 is connected to receive control signals 20 from the vehicle controller 16 via a suitable interface 19. For example, the interface 19 may provide isolation between the high-voltage domain and the low-voltage domain of the system. For this purpose, the interface 19 may comprise an optocoupler.

The control signals 20 may be communicated over a bi-directional serial communication link, such as a CAN bus. Such serial communication is then preferably configured as a safety critical "black channel", e.g. provided with a communication protocol capable of detecting any fault condition.

Alternatively, the control signal 20 is a unidirectional communication of a (binary) discharge signal. In this case it may be useful to provide the interface 19 with circuitry for increasing the reliability of such simple control signals. This will be discussed in detail below with reference to fig. 3.

In use, the controller 13 will receive a control signal 20 from the vehicle controller 16 and provide an activation signal to the switch 12 in response to the signal, thereby causing the switch to enter a closed (conducting) state. Typically, the discharge signal is normally present (high) unless there is a fault condition or damage, at which the discharge signal is absent (low). Thus, in the absence of a discharge signal, the controller 13 provides an activation signal to the switch 12.

When the switch 12 is closed, a constant and predefined current will be drawn by the current source 11 and energy will be dissipated in the current source 11 and any additional dissipation elements 14.

Fig. 2 shows a more detailed embodiment of the active discharge circuit of fig. 1. In this case, the switch 12 is realized by a field effect transistor 21, the source of which field effect transistor 21 is connected (via a resistor 18) to the negative DC line 4 and the drain thereof is connected to the current source 11. The gate of transistor 21 is connected to receive an activation signal from controller 13.

The current source 11 here comprises a field effect transistor 24, the source of which field effect transistor 24 is connected (via a resistor 25) to the drain of the transistor 21 and the gate of which is connected to the cathode of a zener diode 26, the anode of the zener diode 26 being connected to the negative power line 4.

Here, dissipating element 14 is realized by a set (one or more) of field effect transistors 27, which set 27 is connected source to drain between the drain of transistor 24 and positive power line 3. Furthermore, the circuit comprises a set of resistors 28 connected in series between the positive line 3 and the cathode of the diode 26. Each resistor 28 is connected between the gates of adjacent transistors 27 so as to form a string of interconnected resistors 28 and transistors 27.

In use, the voltage across the DC link will be divided by the resistors 28 into a set of intermediate voltages, one intermediate voltage across each resistor 28. Each transistor gate will be subjected to such an intermediate voltage that each transistor 27 is kept in a conducting state.

In addition, there will be a small current i_BiasingThrough the series resistor 28 and finally through the zener diode 26. This current will serve as a bias current to maintain a constant voltage (e.g., about 15V) across the zener diode 26. This constant voltage will be applied to the gate of transistor 24, thereby defining a particular operating state of transistor 24. The bias current will gradually decrease as the charge of the link capacitor 5 discharges. However, as long as the voltage across the capacitor 5 is sufficiently large, the voltage across the zener diode 26 will remain substantially constant. For example, if there are six resistors 28 (as shown in FIG. 2), each having a resistance of 270kOhm, the bias current is about 0.5mA for a capacitor voltage of 800V.

When transistor 21 receives an activation signal from controller 13, transistor 21 will enter a conductive state, allowing current i_{Discharge of electricity}From drain to source. The magnitude of this discharge current will be defined by the state of transistor 24, which is defined by the voltage across zener diode 26 and the voltage developed across resistor 25.

As current flows through the set of transistors 27, energy will be dissipated in each transistor 27 (and also in transistor 24) to gradually discharge capacitor 5. The discharge current through transistors 27, 24 will be constant and the voltage drop across capacitor 5 will be linear as long as the activation signal is present.

The voltage across the zener diode 26, and hence the current drawn by the transistor 24, will remain substantially constant here until the voltage across the capacitor is below 60V, which is a regulatory requirement. In practice, the discharge circuit will continue to be effective below 60V, but then with a slightly lower discharge current, because the voltage across the zener diode 26 will be smaller when the bias current is very small. Eventually, the remaining voltage across the link capacitor 5 will be too small to keep the transistors 24, 27 in their conductive state, and the discharge circuit 10 will be disabled.

In an alternative embodiment, the current sources 11, 24 are configured to draw a larger discharge current than the

dissipative elements

14, 28 can withstand at full capacitor voltage (e.g., 800V). The controller 13 is then configured to provide an intermittent (pulsed) activation signal, starting with a relatively low duty cycle, and then increasing the duty cycle (eventually reaching a permanently on state). With sufficient duty cycle control, the current sources 11, 24 can be controlled to draw increasing average currents so that the dissipated power is constant. The energy discharge of the DC link capacitor 5 will then be linear (and the voltage drop will be exponential).

In yet another embodiment, the controller 13 is again configured to provide an intermittent (pulsed) activation signal, now with a decreasing duty cycle. With sufficient duty cycle control, the current sources 11, 24 can be controlled to draw a decreasing average current such that the voltage across the DC link capacitor decreases exponentially. This behavior is similar to that of a conventional discharge resistor and may be advantageous if the discharge is aligned with the discharge of other capacitors.

During discharging, the controller 13 may be configured to continuously detect the voltage across the DC link using the voltmeter 15 to verify whether the voltage is decreasing as expected. If no drop in voltage is detected, this is an indication of a fault condition, e.g. the power supply 2 is still connected to the DC link. In this case, the continued activation of the active discharge circuit 10 may lead to a thermal event in the dissipating component 14 (transistor 27), potentially damaging the component or even causing a fire hazard. Thus, the controller is preferably configured to deactivate the switch 12 if it determines that the voltage across the DC link has not decreased as expected.

Referring again to fig. 1 and 2, a resistor 18 may be connected between the switch 12 (transistor 21) and the negative line 4. The voltage across the resistor 18 may then be provided to the controller 13 and act as a current detector. This can be used to achieve a simple functional test. The controller 13 may be configured to provide a short (millisecond) activation signal to the switch 12 and then verify whether the detected current is as expected.

In either case, i.e., if the voltage is not reduced as expected or if the functional test fails, the active discharge circuit may be disabled. In this case, the high resistance discharge resistor 6 will provide a "safe state" ensuring that the link capacitor 5 will be discharged (albeit slowly). When the controller 13 communicates with the vehicle controller 16 over a bi-directional communication link (see above), the controller may also communicate to the vehicle controller 16 that the active discharge circuit has been disabled.

By implementing the safety function described above, the active discharge circuit can be designed to comply with relevant safety regulations, such as ISO 26262 and IEC 13849.

Fig. 3 shows an example of the interface 19 in the case where the communication 20 is a discharge signal from the vehicle controller 16. The interface 19 here comprises an optical switch 33, which optical switch 33 comprises an LED 34 and a phototransistor 35. The light dependent resistor 35 is connected between the operating voltages and via a resistor 36 to the negative line 4. The anode of LED 34 is connected to discharge signal 20 and the cathode of LED 34 is connected to the drain of transistor 38. Transistor 38 has a source connected to ground and a gate connected to a pulse signal 40 from a pulse generator 39.

In use, when both the discharge signal 20 and the pulse signal 40 are active, the phototransistor 35 will provide a pulse signal output 37. If either the discharge signal 20 or the pulse signal 40 is not present, there will be no pulse output 37.

Turning to the controller 13, which is shown in greater detail in fig. 3, it includes a processing circuit 41 for verifying that the output from the phototransistor 35 is a pulsed signal. Circuitry 41 may be configured to verify voltage levels, pulse frequency, pulse duration, pulse pattern, or combinations thereof. The processing circuit 41 will provide an activation signal output if and only if the pulse discharge signal cannot be verified. In other words, if the discharge signal 20 is not present, or if the power supply to the pulse generator 39 is terminated, the controller 13 will activate the switch 12 to discharge the capacitor 5.

Referring again to fig. 1 and 2, in the illustrated embodiment, the voltage at the source of transistor 24 is used to drive controller 13. Specifically, here the voltage is applied to the voltage regulating circuit 17, and is from the output (V) of the voltage regulator 17_{Operation of}) QuiltFor powering the controller 13. In the example shown, the voltage across zener diode 26 is approximately 15V, and the voltage between the gate and source of transistor 24 (V)_gs) About 3V, so that the voltage V at the source of transistor 21 is_{Drive the}About 12V. The output from the voltage regulator 17 may typically be 2-5V.

When voltage is used to power the controller, there will be a small reactive current through transistor 27. This reactive current will cause losses and should therefore not be larger than necessary. In particular, the reactive current should be significantly smaller than the discharge current. Typically, the reactive current will be at most one fifth of the discharge current and may be one tenth.

As an example, for a relatively small link capacitor 5, a discharge power of 4W may be sufficient, which for a link voltage of 800V corresponds to a discharge current of 5mA (═ 4W/800V). If the reactive current is 1mA, it will result in a loss of 0.8W (═ 800V × 1mA), which is acceptable.

Note that switch 12 (shown in fig. 2 as transistor 21) may alternatively be implemented by suitable circuitry in controller 13. For example, the controller 13 may be configured to connect the operating voltage from the voltage regulator to ground via a suitable resistor (not shown).

The person skilled in the art realizes that the present invention by no means is limited to the preferred embodiments described above. On the contrary, many modifications and variations are possible within the scope of the appended claims. For example, other types of transistors may be used in place of the field effect transistors shown. Furthermore, the given voltage and current levels are merely exemplary.

Claims

1. An active discharge circuit (10) for an electric vehicle inverter (1), intended to be connected in parallel with a DC link capacitor (5) connected between a positive line (3) and a negative line (4) of a DC power link, and configured to discharge the DC link capacitor (5) in less than seven seconds, wherein the active discharge circuit comprises:

a dissipation current source (11; 24),

a switch (12; 21), the switch (12; 21) being connected in series with the current source between the DC lines, an

A controller (13) connected to the switch and arranged to apply an activation signal in dependence on a control signal (20) from a vehicle control system (16), the activation signal placing the switch in a conducting state,

wherein the current source is configured to draw a discharge current and dissipate any energy stored in the DC link capacitor when the switch is in the on state,

such that the voltage across the DC link capacitor will decrease linearly when the switch is in the on state.

2. The active discharge circuit of claim 1, further comprising a set of active components (14; 27) connected in series between the positive line and the current source.

3. The active discharge circuit of claim 2, wherein the active component is a source-to-drain connected transistor (27).

4. The active discharge circuit of claim 3, further comprising a set of resistors (28) connected in parallel with the DC link capacitor to divide the voltage across the DC link capacitor into a set of intermediate voltages, each intermediate voltage connected to the gate of one of the field effect transistors.

5. Active discharge circuit according to one of the preceding claims, wherein the dissipation current source (11; 24) comprises a transistor and a voltage regulator (26) connected between the gate of the field effect transistor and the negative DC line.

6. The active discharge circuit according to claim 5, wherein the drain of the transistor (24) is connected to the positive line (3) without any intermediate resistive load.

7. Active discharge circuit according to one of the preceding claims, wherein when the switch (12; 21) is in a non-conducting state, a reactive current is allowed to flow through the current source, the reactive current being substantially smaller than the discharge current.

8. Active discharge circuit according to claim 6, wherein the reactive current (i) is_{Reactive power}) At most one fifth of the discharge current and preferably at most one tenth of the discharge current.

9. Active discharge circuit according to claim 7 or 8, wherein the reactive current is used to power the controller (13).

10. The active discharge circuit of any of the preceding claims, wherein the switch comprises a transistor (21), the transistor (21) having a drain connected to the current source, a source connected to the negative DC line, and a gate connected to receive the activation signal.

11. The active discharge circuit according to any of the preceding claims, wherein the controller (13) is configured to apply a stable activation signal, thereby controlling the current source to draw a constant discharge current such that the voltage drop across the DC-link capacitor is linear.

12. The active discharge circuit according to any of claims 1-9, wherein the controller (13) is configured to apply intermittent activation signals with increasing duty cycle, thereby controlling the current source to draw increasing average current to dissipate constant power such that the voltage drop across the DC link capacitor increases exponentially.

13. Active discharge circuit according to any of the claims 1-10, wherein the controller (13) is configured to provide an intermittent activation signal with a decreasing duty cycle, thereby controlling the current source to draw a decreasing average current such that the voltage across the DC-link capacitor decreases exponentially.

14. Active discharge circuit according to one of the preceding claims, wherein the control signal (20) communicates on a bidirectional serial communication bus.

15. Active discharge circuit according to one of the claims 1 to 13, wherein the control signal (20) is a discharge request signal and the controller (13) is configured to apply the activation signal in the absence of the discharge request signal.

16. The active discharge circuit of claim 15, further comprising circuitry (38, 39) for pulsing the discharge request signal (20) to generate a pulsed discharge signal (37), and wherein the controller is configured to validate the pulsed discharge signal (37) and provide the activation signal when the validation is unsuccessful.

17. Active discharge circuit according to any of the preceding claims, wherein the switch (12; 21) is connected to the negative line (4) via a resistor (18) and the controller (13) is connected to detect the voltage over the resistor (18).

18. Active discharge circuit according to any of the preceding claims, further comprising a voltmeter (15), said voltmeter (15) being connected to detect a link voltage between said DC lines (3, 4), and wherein said controller (13) is connected to receive an indication of said link voltage from said voltmeter and to determine whether said link voltage has dropped correctly, and to cause said switch (12) to enter a non-conducting state when it is determined that said link voltage has not dropped correctly.